Solutions to Atiyah-Macdonald, Chapter 1

Dave Karpuk
May 19, 2010
Exercise 1. Let x be a nilpotent element of a ring A. Show that 1 +x is a unit of A. Deduce that
the sum of a nilpotent element and a unit is a unit.
Proof. Note that x is nilpotent i x is nilpotent, so we may replace 1 + x with 1 x. For all
n 1, we have
1 +x
n
= (1 x)(1 +x + +x
n1
).
Thus if x
n
= 0, the element 1 +x + +x
n1
is the inverse of 1 x.
If u is any unit and u
1
its inverse, then u
1
x is nilpotent, so 1 + u
1
x is a unit. Since the
units form a group under multiplication, we have that
u +x = u(1 +u
1
x)
is a unit.
Exercise 2. Let A be a ring and let A[x] be the ring of polynomials in an indeterminate x, with
coecients in A. Let f = a
0
+a
1
x + +a
n
x
n
A[x]. Prove that
(i) f is a unit in A[x] a
0
is a unit in A and a
1
, . . . , a
n
are nilpotent.
(ii) f is nilpotent a
0
, a
1
, . . . , a
n
are nilpotent.
(iii) f is a zero-divisor there exists a = 0 in A such that af = 0.
(iv) f is said to be primitive if (a
0
, a
1
, . . . , a
n
) = (1). Prove that if f, g A[x], then fg is
primitive f and g are primitive.
Proof. (i) Suppose that a
0
is a unit and that the other a
i
are all nilpotent. Then clearly a
i
x
i
is
nilpotent, so a
1
x+. . . +a
n
x
n
is nilpotent. As a
0
is a unit, it follows that f = a
0
+a
1
x+. . . +a
n
x
n
is a unit by exercise 1.
Conversely, suppose that f is a unit. Then we can nd some g = b
0
+ . . . + b
m
x
m
such that
fg = 1. This immediately implies that a
0
is a unit, the hard part is proving the other a
i
are
nilpotent. The equation fg = 1 implies that
0 = a
n
b
m
0 = a
n
b
m1
+a
n1
b
m
.
.
.
0 = a
n
b
0
+a
n1
b
1
+. . . +a
0
b
n
(set b
j
= 0 if j > m)
Multiplying the second equation from the top by a
n
gives 0 = a
2
n
b
m1
. Multiplying the third
equation by a
2
n
gives a
3
n
b
m2
= 0. Continuing in this fashion, we see that there is some power a
j
n
of a
n
we can multiply the last equation by to obtain
0 = a
j+1
n
b
0
.
1
Multiplying both sides of this equation by b
1
0
(which exists) shows that a
n
is nilpotent.
Now a
n
x
n
is nilpotent, hence f a
n
x
n
, the sum of a unit and a nilpotent element, is a unit.
But f a
n
x
n
has degree less than that of f, so we are done by induction on the degree of f.
(ii) If all of the coecients of f are nilpotent, then f is nilpotent as it is the sum of nilpotents.
For the converse, note that f
m
= 0 for some m implies that a
m
n
x
nm
= 0, hence a
n
x
n
is nilpotent.
Thus f a
n
x
n
is the sum of nilpotent elements and hence nilpotent, the result now follows by
induction on the degree of f.
(iii) Let f = a
n
x
n
+. . .+a
0
be a zero-divisor of A[x], such that fg = 0, where g = b
m
x
m
+. . .+b
0
is of minimal degree. Writing out fg immediately yields a
n
b
m
= 0. We have a
n
gf = 0, but since a
n
kills the leading coecient of g, we must have that a
n
g is of degree smaller than g, unless a
n
g = 0.
Now suppose that a
n
g = a
n1
g = = a
ni
g = 0. We have
0 = fg
= a
n
gx
n
+. . . +a
ni
gx
ni
+a
ni1
gx
ni1
+. . . +a
0
g
= a
ni1
gx
ni1
+. . . +a
0
g
The leading coecient of the above polynomial is a
ni1
b
m
= 0. Thus a
ni1
gf = 0, as before.
By induction, we conclude that a
i
g = 0 for all i = 0, . . . , n. In particular, a
i
b
m
= 0 for all i, which
implies that b
m
g = 0.
(iv) Let f and g be as above, and let a = (a
0
, . . . , a
n
), b = (b
0
, . . . , b
m
). Let c = (a
0
b
0
, a
0
b
1
+
b
0
a
1
, . . . , a
n
b
m
) be the corresponding ideal for the product fg. Then the problem becomes to show
that a = (1) = b if and only if c = (1).
For one direction, note that c a and c b, so c = 1 implies a = (1) = b trivially.
For the other direction, suppose that a = (1) = b.
Exercise 3. Generalize the results of Exercise 2 to a polynomial ring A[x
1
, . . . , x
n
] in several
indeterminates.
Proof.
Exercise 4. In the ring A[x], the Jacobson radical is equal to the nilradical.
Proof. We need only to show that the Jacobson radical is contained in the nilradical, as the other
inclusion holds for all rings.
Let f = a
0
+ +a
n
x
n
be in the Jacobson radical. By Proposition 1.9, 1 fg is a unit for all
g A[x]. Setting g = 1, we obtain that 1 +f is a unit. Then by Exercise 2(i), a
i
is nilpotent for
i 1. Setting g = (1 +x), we obtain that a
0
+a
1
is nilpotent (because it is the coecient of x in
1 fg), hence a
0
is nilpotent. Thus all of the coecients of f are nilpotent, and by Exercise 2(ii),
this is equivalent to f being nilpotent.
Exercise 5. Let A be a ring and let A[[x]] be the ring of formal power series f =

n=0
a
n
x
n
with
coecients in A. Show that
(i) f is a unit in A[[x]] a
0
is a unit in A.
(ii) If f is nilpotent, then a
n
is nilpotent for all n 0. Is the converse true?
(iii) f belongs to the Jacobson radical of A[[x]] a
0
belongs to the Jacobson radical of A.
2
(iv) The contraction of a maximal ideal m of A[[x]] is a maximal ideal of A, and m is generated
by m
c
and x.
(v) Every prime ideal of A is the contraction of a prime ideal of A[[x]].
Proof. (i) Let g = b
0
+b
1
x + . If fg = 1, then a
0
b
0
= 1 so a
0
is a unit.
Conversely, suppose that f is a unit. For any g, we have
fg =

k=0
_
_

i+j=k
a
i
b
j
_
_
x
k
so for g to be an inverse to f, we must be able to solve the equations
a
0
b
0
= 1,

i+j=k
a
i
b
j
= 0 for k 1.
The rst equation we can solve because a
0
is a unit in A. By induction on k, we can solve the rest
of the equations by setting
b
k
= a
1
0
(a
1
b
k1
+ +a
k
b
0
).
Thus f is a unit in A[[x]].
(ii) Suppose that f
N
0
= 0. Then a
N
0
0
= 0, so a
0
is nilpotent. As the dierence of two nilpotent
elements is nilpotent, we have that
f a
0
= a
1
x +a
2
x
2
+
is nilpotent. Pick N
1
such that (f a
0
)
N
1
= 0. Then a
N
1
1
= 0, so a
1
is nilpotent, which implies
that
f a
0
a
1
x = a
2
x
2
+
is nilpotent. Continuing in this fashion, we see that for all n, there is some N
n
such that
(f a
0
a
1
x a
n1
x
n1
)
Nn
= 0
which implies that a
Nn
n
= 0.
To see that the converse fails, let y
n
be a collection of indeterminates over Z, indexed by the
natural numbers. Let
A = Z[y
1
, y
2
, y
3
. . .]/(y
1
, y
2
2
, y
3
3
, . . .).
Let f =

n=1
y
n
x A[[x]]. Then a
n
= y
n
is nilpotent for all n, but there is no N such that f
N
= 0
(this would establish some non-trivial relation amongst the y
n
for n > N).
(iii) A power series f is in the Jacobson radical of A[[x]] if and only if for all g A[[x]], 1 fg
is a unit in A[[x]]. By part (i), this is equivalent to 1 a
0
b
0
being a unit for all b
0
A, which is
equivalent to a
0
being in the Jacobson radical of A.
(iv) Let m be any maximal ideal in A[[x]]. Suppose that x / m. Since A[[x]]/m is a eld, there
exists f A[[x]] such that 1 fx (m). Then 1 fx m, but 1 fx is a unit by part (i). Thus
x m for all maximal ideals m of A[[x]].
3
We must show that A/m
c
is a eld. Let a A be any non-zero element. We must show that it
has a multiplicative inverse modulo m
c
. Since A[[x]]/m is a eld, there exists f = a
0
+ a
1
x +
such that af 1 (m). Since x m, we have af aa
0
(m), and thus aa
0
1 (m
c
).
To prove the second part, note that we have the obvious inclusion (m
c
, x) m. To see the
reverse inclusion, note that if f = a
0
+ a
1
x + m, then a
0
m because x m. Hence
a
0
m A = m
c
, so that f can be written as a sum of elements of m
c
and (x). This proves the
desired equality.
(v) Let p be a prime ideal of A. Then the set p[[x]] of elements of A[[x]] with coecients in p is an
ideal of A[[x]]. It is prime because it is the kernel of the natural homomorphism A[[x]] (A/p)[[x]],
given by reducing all of the coecients modulo p. The ring (A/p)[[x]] is an integral domain because
A/p is. Clearly p = A p[[x]] = p[[x]]
c
.
Exercise 6. A ring A is such that every ideal not contained in the nilradical contains a non-zero
idempotent (that is, an element e such that e
2
= e = 0). Prove that the nilradical and the Jacobson
radical are equal.
Proof. Let N be the nilradical, J be the Jacobson radical, and suppose that N J. Then there
exists a non-zero idempotent e J. We have e(1 e) = 0, but 1 e is a unit by Proposition 1.9.
Hence e = 0, a contradiction.
Exercise 7. Let A be a ring in which every element x satises x
n
= x for some n > 1 (depending
on x). Show that every prime ideal in A is maximal.
Proof. Let x Ap, and let x be the corresponding non-zero element in A/p. Then x( x
n1
1) = 0
in A/p, an integral domain. Thus x
n1
= 1, so x has multiplicative inverse x
n2
. Therefore A/p is
a eld, which says that p is maximal.
Exercise 8. Let A be a ring = 0. Show that the set of prime ideals of A has minimal elements
with respect to inclusion.
Proof. Let S be the set of all prime ideals of A, ordered with respect to reverse inclusion. The
set S is non-empty by Theorem 1.3. If {p
i
} is a chain in S, then it has an upper bound, namely

i
p
i
. By Zorns Lemma, S has maximal elements, which are prime ideals minimal with respect to
inclusion.
Exercise 9. Let a be an ideal = (1) in a ring A. Show that a = r(a) a is an intersection of
prime ideals.
Proof. By Proposition 1.8, r(a) is equal to the intersection of all of the prime ideals containing a.
Thus a = r(a) if and only if a is equal to the intersection of all prime ideals containing it.
Exercise 10. Let A be a ring, N its nilradical. Show that the following are equivalent:
(i) A has exactly one prime ideal;
(ii) every element of A is either a unit or nilpotent;
(iii) A/N is a eld.
4
Proof. (i)(ii) If A has only one prime ideal, then it must be N, since N is the intersection of all
prime ideals of A. Hence the units of A are exactly A N, which are exactly the non-nilpotent
elements.
(ii)(iii) Since everything in AN is a unit, we have that A/N is a eld.
(iii)(i) If A/N is a eld, then its only prime ideal is (0). Hence the only prime ideal containing
N in A is N itself. But N is the intersection of all prime ideals of A, so A must have only one
prime ideal, namely N.
Exercise 11. A ring A is Boolean if x
2
= x for all x A. In a Boolean ring A, show that
(i) 2x = 0 for all x A;
(ii) every prime ideal p is maximal, and A/p is a eld with two elements;
(iii) every nitely generated ideal in A is principal.
Proof. (i) Let x A, then
2x = x +x = (x +x)
2
= x
2
+ 2x
2
+x
2
= x + 2x +x = 2x + 2x
and so 2x = 0.
(ii) By Exercise 7, every prime ideal of A is maximal. Now suppose that x A/p is non-zero.
Then x
2
= x implies that x = 1, so that the only non-zero element of A/p is the multiplicative
identity. Thus A/p only has two elements.
(iii) By induction, it suces to prove that every ideal generated by two elements is principal.
Let a = (x, y) be such an ideal in A. Then x(x + y xy) = x
2
+ xy x
2
y = x + xy xy = x.
Similarly y(x +y xy) = y, so (x, y) = (x +y xy), proving that this ideal is principal.
Exercise 12. A local ring contains no idempotent = 0, 1.
Proof. Let A be a local ring with maximal ideal m. Suppose that we had an idempotent e = 0, 1.
Then e is a zero-divisor and thus cannot be a unit, so e m. So 1 e is a unit by Proposition
1.6(ii). But 1 e is also a zero-divisor, hence we have a contradiction.
Exercise 13. Let K be a eld and let be the set of all irreducible monic polynomials f in
one indeterminate with coecients in K. Let A be the polynomial ring over K generated by
indeterminates x
f
, one for each f . Let a be the ideal of A generated by the polynomials f(x
f
)
for all f . Show that a = (1).
Let m be a maximal ideal of A containing a, and let K
1
= A/m. Then K
1
is an extension eld
of K in which each f has a root. Repeat the construction with K
1
in place of K, obtaining
a eld K
2
, and so on. Let L =

n=1
K
n
. Then L is a eld in which each f splits completely
into linear factors. Let

K be the set of all elements on L which are algebraic over K. Then

K is
an algebraic closure of K.
Proof.
Exercise 14. In a ring A, let be the set of all ideals in which every element is a zero-divisor.
Show that the set has maximal elements and that every maximal element of is a prime ideal.
Hence the set of zero-divisors in A is a union of prime ideals.
5
Proof. Clearly (0) , so is non-empty. Let (a
i
)
iI
be a chain in . The union a of the chain is
still an ideal, and consists only of zero-divisors. Thus a , so by Zorns Lemma has maximal
elements.
Let a be a maximal element of , and suppose that xy a with y / a. Then a (a, y) so y is
not a zero-divisor by the maximality of a. But xy is a zero-divisor, so there exists some z A such
that zxy = 0. As y is not a zero-divisor we must have that xz = 0, so that x is a zero-divisor and
thus is contained in a.
Exercise 15. Let A be a ring and let X be the set of all prime ideals of A. For each subset E of
A, let V (E) denote the set of all prime ideals of A which contain E. Prove that
(i) if a is the ideal generated by E, then V (E) = V (a) = V (r(a)).
(ii) V (0) = X, V (1) = .
(iii) if (E
i
)
iI
is any family of subsets of A, then
V
_
_
iI
E
i
_
=

iI
V (E
i
).
(iv) V (a b) = V (ab) = V (a) V (b) for any ideals a, b of A.
Proof. (i) For the rst equality note that E p if and only if a p, so V (E) = V (a). For the
second equality, we use the fact that r(a) is the intersection of all prime ideals containing a, so
a p if and only if r(a) p. Thus V (a) = V (r(a)).
(ii) 0 p for all p, so V (0) = X. The element 1 is not contained in any prime ideal, since prime
ideals are proper by denition. Thus V (1) = .
(iii) We have p V (

i
E
i
) p

i
E
i
p E
i
for all i p V (E
i
) for all i
p

i
V (E
i
)
(iv) By Exercise 1.13 in the text we have r(ab) = r(a b), so by part (i) of this exercise we
have V (ab) = V (a b).
Now suppose that p V (a b), so that a b p. By Proposition 1.11(ii) it follows that either
a p or b p. Therefore p V (a) V (b). Therefore V (a b) V (a) V (b). To see the reverse
inclusion, let p V (a) V (b) so that either a p or b p. In any case, a b p, so p V (a b).
Therefore V (a b) = V (a) V (b).
Exercise 16. Draw pictures of Spec (Z), Spec (R), Spec (C[x]), Spec (R[x]), Spec (Z[x]).
Proof. Spec (Z) is a line, with one closed point for each prime number in Z, and one non-closed
point corresponding to the ideal (0).
As R is a eld, Spec (R) consists of a single point, corresponding to the ideal (0).
The maximal ideals of the ring C[x] are all of the form (x ) for C (because C is
algebraically closed). The only non-maximal prime ideal of C[x] is (0). Thus Spec (C[x]) is the
complex plane together with a single non-closed point, corresponding to (0).
There are two types of maximal ideals in R[x], those generated by a polynomial of degree one,
and those generated by a polynomial of degree two. The degree one maximal ideals are all of the
form (x ) for R. The maximal ideals of degree two correspond to complex conjugate pairs
of elements of C R. There is only one non-maximal prime ideal, (0). Therefore Spec (R[x]) is
the upper-half plane, together with a single non-closed point.
For a drawing of Spec (Z[x]), see The Geometry of Schemes by Eisenbud and Harris.
6
Exercise 17. For each f A, let X
f
denote the complement of V (f) in X = Spec (A). The sets
X
f
are open. Show that they form a basis of open sets for the Zariski topology, and that
(i) X
f
X
g
= X
fg
;
(ii) X
f
= f is nilpotent;
(iii) X
f
= X f is a unit;
(iv) X
f
= X
g
r((f)) = r((g));
(v) X is quasi-compact.
(vi) More generally, each X
f
is quasi-compact.
(vii) An open subset of X is quasi-compact if and only if it is the nite union of sets X
f
.
Proof. To see that these open sets form a basis for the topology on Spec (A), let U = X V (a) be
any open set in Spec (A). Then
U = {p|a p}
= {p| there exists f a, f / p}
=
_
fa
{p|f / p}
=
_
fa
X
f
so every open set can be written as a union of the sets X
f
.
(i) From the denition of prime ideal, we see that p X
f
X
g
f / p and g / p
fg / p p X
fg
.
(ii) X
f
= f p for all p f

p f is nilpotent.
(iii) X
f
= X f / p for any p f is a unit.
(iv) If X
f
= X
g
, then V (f) = V (g). Thus f and g belong to exactly the same prime ideals,
hence r((f)) = r((g)). Conversely, if r((f)) = r((g)), then V (f) = V (g) by Exercise 15(i). This
implies that X
f
= X
g
.
(v) Suppose we have an open cover of X by basic sets X
f
i
, i I. Then for all prime ideals
p A, there exists i such that f
i
/ p. Therefore the f
i
generate the unit ideal, so we have an
equation
n

i=1
a
i
f
i
= 1
for some a
1
, . . . , a
n
A. Thus f
1
, . . . , f
n
generate the unit ideal, so does not exist a prime ideal p
such that f
1
, . . . , f
n
p. Therefore
X = X
f
1
X
fn
.
Since every open cover by basic open sets has a nite subcover, it follows that X is quasi-compact
since the basic open sets form a basis for the topology of X.
7
(vi) Suppose X
g
=

i
X
f
i
. Then for all primes p such that g / p, there exists i such that f
i
/ p.
Let a be the ideal generated by all of the f
i
. Then if a p for some p, we have that g p, and
hence g r(a). We can therefore nd an m such that g
m
a. So we have an equation of the form
g
m
=
n

i=1
a
i
f
i
for some f
1
, . . . , f
n
. If f
1
, . . . , f
n
all belong to some prime ideal p, we have g
m
p and therefore
g p because p is prime. Equivalently, g / p for some p f
i
/ p for some i = 1, . . . , n. This
implies that X
g
= X
f
1
X
fn
, and thus X
g
is quasi-compact.
(vii) This follows directly from the fact that every open set of X can be written as a union of
X
f
s.
Exercise 18. Let x X = Spec (A) and let p
x
be the corresponding prime ideal. Show that
(i) the set {x} is closed in X p
x
is maximal;
(ii) {x} = V (p
x
);
(iii) y {x} p
x
p
y
;
(iv) X is a T
0
-space.
Proof. (i) {x} is closed {x} = V (a) for some a p
x
is the only prime ideal containing
some ideal a p
x
is maximal, because every a is contained in some maximal ideal.
(ii) We have
{x} =

xV (a)
V (a) =

apx
V (a) = V
_
_
apx
a
_
= V (p
x
).
(iii) y {x} y V (p
x
) p
x
p
y
(iv) Let x, y X be distinct. Suppose that every neighborhood of x contains y. Since the sets
X
f
form a basis for the topology of Spec (A), this is equivalent to (x X
f
y X
f
for all f A).
But this is equivalent to (f / p
x
f / p
y
for all f A), which is the same as p
y
p
x
. Thus is
every neighborhood of y contained x, we would have p
x
p
y
, forcing p
x
= p
y
. But this says x = y,
contradicting the fact that x and y are distinct. Thus there exists some neighborhood of y not
containing x, proving that X is T
0
.
Exercise 19. Show that Spec (A) is irreducible if and only if the nilradical of A is a prime ideal.
Proof. Suppose that the nilradical of A is prime, and let X
f
, X
g
be two basic open sets in Spec (A).
If X
f
X
g
= , then X
fg
= by Exercise 17(i). Thus fg is nilpotent, so either f or g is nilpotent,
which says that either X
f
= or X
g
= by Exercise 17(ii). Therefore any two non-empty basic
open sets intersect non-trivially. Because these sets form a basis for the topology of Spec (A), it
follows that Spec (A) is irreducible.
Conversely, suppose that the nilradical is not prime. Then there exists non-zero f, g A such
that fg is nilpotent but neither f nor g is nilpotent. By Exercise 17(i) and (ii), it follows that
neither X
f
nor X
g
is empty, but X
f
X
g
= X
fg
= . Thus Spec (A) is not irreducible.
Exercise 20. Let X be a topological space.
(i) If Y is an irreducible subspace of X, then the closure

Y of Y in X is irreducible.
(ii) Every irreducible subspace of X is contained in a maximal irreducible subspace.
8
(iii) The maximal irreducible subspaces of X are closed and cover X. What are the irreducible
components of a Hausdor space?
(iv) If A is a ring and X = Spec (A), then the irreducible components of X are the closed sets
V (p), where p is a minimal prime ideal of A.
Proof. (i) Let U

Y be open and non-empty. Then UY is open in Y and thus dense in Y , because
Y is irreducible. Hence the closure of U Y in

Y is dense in

Y , but this is the same as the clo-
sure of U in

Y . Therefore every non-empty open set in

Y is dense, which shows that

Y is irreducible.
(ii) Let Y be an irreducible subspace of X, and let be the set of all irreducible subsets of
X containing Y , partially ordered with respect to inclusion. Then Y , so is non-empty. If
(Z
i
)
iI
is a chain in , let Z =

i
Z
i
. We wish to show that Z is irreducible.
Let U, V Z be open and non-empty. We must show that they intersect. As Z =

i
Z
i
and
(Z
i
)
iI
is a chain, there must exist an i such that U Z
i
= and V Z
i
= . But Z
i
is irreducible,
and both U Z
i
and V Z
i
are open in Z
i
. Hence they intersect non-trivially, which says that U
and V intersect non-trivially. Therefore Z is irreducible.
Clearly Y Z, so Z and thus has maximal elements by Zorns Lemma. The maximal
elements of are exactly the maximal irreducible subspaces of X.
(iii) By part (i) of this exercise, the maximal irreducible subspaces of X are closed. For any
topological space X, a singleton {x} is vacuously irreducible and thus contained in a maximal
irreducible subspace. Hence the maximal irreducible subspaces cover X.
Now let X be Hausdor and Y an irreducible subspace. If there are distinct points x, y Y ,
then we can separate them with disjoint open neighborhoods U
x
, U
y
in X. Then U
x
Y, U
y
Y
are open subsets of Y which are disjoint, which contradicts the irreducibility of Y . Therefore Y
consists of just a single point.
(iv) We will rst prove the following statement: If V (a) is a closed subset of X = Spec (A),
then V (a) is irreducible r(a) is prime.
() Suppose that V (a) is irreducible. Then for any two basic open sets X
f
, X
g
of X, we have
that X
fg
V (a) = X
f
V (a) = or X
g
V (a) = . But X
fg
V (a) = fg r(a), and
similarly for X
f
and X
g
. Hence fg r(a) f r(a) or g r(a), which says exactly that r(a) is
prime.
() Suppose that r(a) is prime. As V (r(a)) = V (a), we may replace r(a) with a and suppose
that a is prime. Suppose we can write V (a) = V (a
1
) V (a
2
) = V (a
1
a
2
). Then a
1
a
2
a, so
either a
1
a or a
2
a by Proposition 1.11. WOLOG a
1
a. Then V (a) V (a
1
) and we must
therefore have equality. This proves that V (a) is irreducible.
To nish the exercise, note that V (a) V (b) r(b) r(a). Thus V (a) = V (r(a)) is an
irreducible component of X if and only if r(a) is a minimal prime ideal.
Exercise 21. Let : A B be a ring homomorphism. Let X = Spec (A) and Y = Spec (B),
and let

: Y X be the induced map. Show that

(i) If f A then
1
(X
f
) = Y
(f)
, and hence that

is continuous.
(ii) If a is an ideal of A, then
1
(V (a)) = V (a
e
).
(iii) If b is an ideal of B, then

(V (b)) = V (b
c
).
(iv) If is surjective, then

is a homeomorphism of Y onto the closed subset V (ker()) of X.

9
(v) If is injective, then

(Y ) is dense in X. More precisely,

(Y ) is dense in X
ker() N.
(vi) Let : B C be another ring homomorphism. Then ( )

.
(vii) Let A be an integral domain with just one non-zero prime ideal p, and let K be the eld
of fractions of A. Let B = (A/p) K. Dene : A B by (x) = ( x, x), where x is the image of
x in A/p. Show that

is bijective but not a homeomorphism.

Proof. (i) We have

1
(X
f
) = {q B|f /
1
(q)}
= {q B|(f) / q}
= Y
(f)
.
Since inverse images of basic open sets are open, it follows that the inverse image of every open set
is open, and thus

is continuous.
(ii) We have

1
(V (a)) = {q B|

(q) V (a)}
= {q B|a
1
(q)}
= {q B|(a) q}
= {q B|a
e
q}
= V (a
e
).
(iii) We have

(V (b)) = {p A|

(V (b)) p}
= {p A|
1
(q) p for all q b}
= {p A|
1
(b) p}
= V (b
c
).
(iv) If is surjective, then we can identify with a natural projection : A A/a. Then
ker() = a. According to Proposition 1.1 and the fact that
1
of a prime ideal is prime, we have
that

is a bijection:
{prime ideals of A/a} {prime ideals of A containing a}.
But this is exactly a bijection

: Y V (a). By part (i),

is continuous and thus a homeomor-

phism.
(v) By part (iii), we have

(Y ) =

(V (0)) = V (
1
(0)) = V (ker())
so

(Y ) is dense V (ker()) = X. But this happens exactly if ker() p for all p, which is
equivalent to ker() N.
10
(vi) This follows from the obvious fact that for any prime ideal p C, we have ( )
1
(p) =

1
(
1
(p)).
(vii) The prime ideals of B are q
1
= (0) K and q
2
= A/p (0). The prime ideals of A are
(0) and p. Clearly,

(q
1
) = p and

(q
2
) = (0), so

is bijective. Now Spec (B) is the two point

Hausdor space, because both of the prime ideals of B are maximal and hence their corresponding
points are closed. But Spec (A) is non-Hausdor because {(0)} is not closed. Therefore

cannot
be a homeomorphism.
Exercise 22. Let A =

n
i=1
A
i
be the direct product of the rings A
i
. Show that Spec (A) is the
disjoint union of the open (and closed) subspaces X
i
, where X
i
is canonically homeomorphic with
Spec (A
i
).
Conversely, let A be any ring. Show that the following statements are equivalent:
(i) X = Spec (A) is disconnected.
(ii) A

= A
1
A
2
where neither of the rings A
1
, A
2
is the zero ring.
(iii) A contains an idempotent = 0, 1.
Proof. Let
i
: A A
i
be the natural projection. As
i
is surjective, we have by Exercise 21(iv)
that

i
: Spec (A
i
) V (ker(
i
)) is a homeomorphism. Setting X
i
= V (ker(
i
)), we have that X
i
is closed.
To see that Spec (A) is the union of the X
i
, let p A be a prime ideal. Let 1
i
A be the
element that has a 1 in the i
th
spot and 0s elsewhere. As 1 / p, we must have that 1
i
/ p for some
i. If j = i then 1
i
1
j
= 0 p, so 1
j
p for all j = i. This implies that
A
1
A
i1
(0) A
i+1
A
n
p
so p is of the form
A
1
A
i1
p
i
A
i+1
A
n
for some p
i
A
i
. Hence p X
i
.
Clearly the X
i
s are disjoint, so it only remains to prove that they are open. But in the course of
the above proof, we showed that X
i
is equal to the basic open set X
1
i
. Therefore X is the disjoint
union of the clopen subspaces X
i
.
To prove the equivalency of the converse statements, note that we have just proven (ii)(i).
(ii)(iii). For the forward direction, suppose that A = A
1
A
2
. Then the elements (1, 0) and
(0, 1) are idempotents, and neither is equal to 0 or 1. For the backwards direction, let e A be
an idempotent. Then Ae and A(1 e) are both rings, with multiplicative identities e and 1 e,
respectively. The map A Ae given by multiplication by e has kernel A(1 e), and it admits a
section ae a. Thus we have a decomposition A = Ae A(1 e).
(i) (iii) (This follows the argument given in Eisenbuds Commutative Algebra on pgs 85-86)
Let X = X
1

X
2
where X
i
is clopen. Because X
i
is closed, we have X
i
= V (a
i
) for some ideals a
i
of A. By Exercise 15(iii) we have V (a
1
+ a
2
) = X
1
X
2
= , so a
1
+ a
2
= (1). Thus there exists
a
i
a
i
such that a
1
+a
2
= 1.
11
By Exercise 15(iv) we have V (a
1
a
2
) = X
1
X
2
= X, so a
1
a
2
p for all p and thus a
1
a
2
is
contained in the nilradical of A. Thus there exists n such that (a
1
a
2
)
n
= 0. Consider the equation
1 = 1
2n
= (a
1
+a
2
)
2n
= a
2n
1
+ +a
n+1
1
a
n1
2
+a
n
1
a
n
2
+a
n1
1
a
n+1
2
+ +a
2n
2
and let e
1
= a
2n
1
+ + a
n+1
1
a
n1
2
and e
2
= a
n1
1
a
n+1
2
+ + a
2n
2
. We have e
1
+ e
2
= 1, and
every monomial in e
1
e
2
contains a factor of (a
1
a
2
)
n
, thus e
1
e
2
= 0. Therefore e
1
= e
1
(e
1
+ e
2
) =
e
2
1
+e
1
e
2
= e
2
1
, so e
1
and e
2
are idempotent.
Exercise 23. Let A be a Boolean ring, and let X = Spec (A).
(i) For each f A, the set X
f
is both open and closed in X.
(ii) Let f
1
, . . . , f
n
A. Show that X
f
1
X
fn
= X
f
for some f A.
(iii) The sets X
f
are the only subsets of X which are both open and closed.
(iv) X is a compact Hausdor space.
Proof. (i) We must prove that X
f
= V (a) for some a. Because A is Boolean, we have that
f(1 f) = 0. As 0 p for every p, it follows easily that f p 1 f / p. Therefore
X
f
= V (1 f), so X
f
is closed.
(ii) By part (i),
X
f
1
X
fn
= V (1 f
1
) V (1 f
n
)
= V
_
n

i=1
(1 f
i
)
_
= X
1

i
(1f
i
)
(iii) Let Y X be open and closed. Then Y is quasi-compact because X is. Let Y =

i
X
f
i
be an open cover by basic open sets. It has a nite subcover, say Y = X
f
1
X
fn
. By part
(ii), Y = X
f
for some f.
(iv) We only have to show that X is Hausdor. Let x and y be distinct points of X, corre-
sponding to prime ideals p
x
and p
y
. Then there exists f p
x
such that f / p
y
. By an observation
from part (i), we then have 1 f / p
x
and 1 f p
y
. Thus x X
1f
and y X
f
, and these open
sets are disjoint. Therefore X is Hausdor.
12